Microsoft just admitted it made a mistake with its AI strategy. Now it's spending $2.5 billion to fix it. The company on Thursday announced Microsoft Frontier Company, a new operating entity that will embed 6,000 engineers, consultants, and sales staff inside customer organizations to build and run AI systems.
The initiative comes with $2.5 billion in funding from Microsoft and will be led by Rodrigo Kede Lima, previously president of Microsoft Asia. The timing is no coincidence. Amazon committed $1 billion to its own forward-deployed engineering unit two days earlier.
OpenAI and Anthropic both launched similar ventures in May, with Anthropic's effort backed by Goldman Sachs, Blackstone, and Hellman & Friedman in a $1.5 billion deal.
Microsoft is pitching Frontier Co. as something bigger than the "forward deployed engineering" label its rivals have embraced. Judson Althoff, CEO of Microsoft's commercial business, called it "the largest, most capable, outcome-driven engineering organization in the industry."
The subtext is more interesting. Microsoft partly owns OpenAI and spent years tying its AI products exclusively to OpenAI's models.
Satya Nadella has since called that a strategic error. "Three years ago, when we built Copilot, we made a mistake by binding it to OpenAI models only," Althoff told Reuters. "You wanted models to amplify your intelligence and be able to have that sort of swappability."
Frontier Co. will help customers pick and integrate AI tools from Microsoft, OpenAI, Anthropic, or open-source providers. Critically, customers keep the results of that work.
Microsoft won't feed client data into its training models. The company is already working with Unilever, Novo Nordisk, and the London Stock Exchange Group.
Microsoft's urgency has a financial dimension. Its stock is down 21% this year, the worst performance among mega-cap tech companies. The Microsoft 365 Copilot assistant hasn't achieved widespread business adoption, and GitHub Copilot has lost market share to competitors.
Enterprise AI services represent a rare growth area. Microsoft generated about $2.1 billion from enterprise and partner services in the March quarter, up 2.5% from a year earlier. But Frontier Co. faces a credibility question: is this actually new?
Microsoft already runs Industry Solutions Delivery, a large consulting arm, along with FastTrack deployment programs and a $1 billion EY alliance. The company declined to say whether the $2.5 billion is fresh capital or repurposed from existing budgets.
What's clear is that the AI industry is converging on a single thesis: selling models alone doesn't work. Companies buy ChatGPT subscriptions and find the demos don't translate to results. The money is in sending engineers to sit inside client operations, figure out where AI actually helps, and build it in.
Microsoft wants to own that layer. The question is whether 6,000 engineers arriving two years late is an advantage or a handicap.
